Prediction of Sinter Quality Based on Population Genetic Algorithms and Fuzzy Reasoning
Download as PDF
DOI: 10.25236/csam.2019.070
Author(s)
Liu Guangyue, Hu Qinghe, Zhang Shuang
Corresponding Author
Liu Guangyue
Abstract
In view of the characteristics of non-linearity, strong coupling and large time delay in sintering process, the sintering process was analyzed from the point of view of process parameter control, and the evaluation index of sinter performance and its main influencing parameters were determined. On this basis, a sinter quality prediction method based on population genetic algorithm and fuzzy reasoning was proposed. Firstly, the pheromone-based ant colony algorithm is used to generate an optimal solution as the initial population of genetic algorithm, which is conducive to improving the convergence performance. Then, by making full use of random numbers and increasing the number of cycles, the roulette-based selection algorithm in traditional genetic algorithm is improved, so as to ensure the diversity of the next generation population and improve the chance of the best chromosome being selected. The method is applied to the prediction of sinter quality. The experimental results show that this method can predict the iron content of ore by quantifying the sinter quality, which has a positive effect on the beneficiation and output prediction of related enterprises.
Keywords
Population Algorithm, Genetic Algorithm, Fuzzy Reasoning, Sinter Quality Prediction